Mammography is currently the most effective method of screening for breast cancer. The goal of breast cancer screening is the detection of early non-palpable tumors. Although mammography is very sensitive in the detection of cancer, it is not very specific in determining whether mammographic abnormalities are due to benign or malignant disease (Limitations of Mammography in the Identification of Noninfiltrating Carcinoma of the Breast, S. F. Sener, F. C. Candela, M. L. Paige, J. R. Bernstein, D. P. Winchester, Surgery, Gynecology, and Obstetrics, August 1988, 167:135-140). Therefore, a noninvasive method of confirming the malignancy of suspicious mammographic abnormalities would be a major benefit in patient care. In this way, the number of benign excisional biopsies (approximately 75% of all excisional biopsies) can be reduced.
When abnormal mammograms are encountered, the physician's options are limited. For minimally suspicious lesions, short-term repeat examination (four to six month follow-up) is often recommended. This may result in psychological stress for the patient and introduces the possibility of loss in patient follow-up due to scheduling or communication errors. The unlikely possibility of interim tumor growth cannot be definitely ruled out (Breast Cancer: Age-Specific Growth Rates and Screening Strategies, M. Moskowitz, Radiology, October 1986, 161:37-41), especially in patients under fifty.
The role of ultrasound in clarifying the status of a mammographic abnormality is limited to the differentiation of solid masses from benign cysts. If the strict criteria for the ultrasonic appearance of a simple cyst are satisfied, the referring physician may be reassured that the lesion is benign. Unfortunately, the current spatial resolution of ultrasound makes the technique of limited value for lesions significantly smaller than five millimeters.
Doppler ultrasound has been advocated as a means for differentiating benign from malignant masses, but results of clinical trials have been contradictory, and the doppler method has no current clinical role in breast imaging (The Role of US in Breast Imaging, V. P. Jackson, Radiology, November 1990, 177:305-311).
Fine-Needle Aspiration (FNA) of breast masses is a technique whose sensitivity and specificity is operator dependent (Fine-Needle Aspiration Biopsies of Breast Masses, L. Palombini et al., Cancer, Jun. 1, 1988, 61:2273-2277), and has been considered experimental (Discriminating Analysis Uncovers Breast Lesions, D. B. Kopans, Diagnostic Imaging, September 1991, pp. 94-101). Because of its relatively low cost and reduced morbidity associated with surgery and anesthesia, FNA has been suggested as a possible replacement for excisional biopsy. Unfortunately, there is a high (13-50%) rate of insufficient samples when FNA is performed on non-palpable mammographically detected lesions. All of these cases of negative FNAs require excisional biopsy (Fine-Needle Aspiration Cytology in Lieu of Open Biopsy in Management of Primary Breast Cancer, H. J. Wanebo et al., Annals of Surgery, May 1984, 199 (5) pp. 569-579). Further, FNA as a non-imaging diagnostic modality, has the disadvantage that no information is obtained about the physical distribution of the detected tumor. As a cytopathological technique, FNA cannot easily differentiate between cases of marked dysplasia, carcinoma-in-situ, or invasive cancer. Fine-Needle Aspiration is generally not performed for non-palpable breast lesions.
Another option for the referral of a patient with equivocal mammographic anomalies is excisional biopsy of the breast in the area corresponding to the region of mammographic abnormality. The probability of malignancy ranges from 2% for a circumscribed solid mass to almost 90% for a spiculated ill-defined mass (Discriminating Analysis Uncovers Breast Lesions, D. B. Kopans, Diagnostic Imaging, September 1991, pp. 94-101. The true-positive fraction for biopsies obtained as a result of a mammographic screening program is between twenty and thirty percent (Nonpalpable Breast Lesions: Accuracy of Prebiopsy Mammographic Diagnosis, G. Hermann, C. Janus, I. S. Schwartz, B. Krivisky, S. Bier, J. G. Rabinowitz, Radiology, November 1987 165:323-326; R. Brem, personal communication). Excisional biopsy has the additional disadvantage of introducing scarring, which may render interpretation of follow-up mammograms more difficult (Discriminating Analysis Uncovers Breast Lesions, D. B. Kopans, Diagnostic Imaging, September 1991, pp. 94-101). An additional disadvantage to excisional biopsies is that, as a non-imaging modality, the physical distribution of the tumor is poorly described.
It is also possible to use radionuclide imaging to detect cancers. 2-[F-18]-Fluoro-2-deoxy-D-glucose (FDG) is a radioactive analogue of glucose that is taken up preferentially by cancer cells (Primary and Metastatic Breast Carcinoma: Initial Clinical Evaluation with PET with the Radiolabeled Glucose Analogue 2-[F-18]-Fluoro-2-deoxy-D-glucose, R. L. Wahl, R. L. Cody, G. D. Hutchins, E. E. Mudgett, Radiology (1991) 179:765-770). A Fluorine-18 nucleus decays by emitting a positron which is annihilated within a millimeter by an electron. The result of this annihilation is the production of two 511 kev (thousand electron volts) gamma rays that are approximately 180 degrees apart in direction. After a patient has received an intravenous dose of FDG she may be examined with detectors that sense these gamma rays.
Previous detection methods have included imaging with a specially collimated planar gamma camera ([18-F] Fluorodeoxyglucose scintigraphy in diagnosis and follow up of treatment in advanced breast cancer, European Journal of Nuclear Medicine (1989) 15:61-66) and with a whole-body Positron Emission Tomography (PET) scanner (Primary and Metastatic Breast Carcinoma: Initial Clinical Evaluation with PET with the Radiolabeled Glucose Analogue 2-[F-18]-Fluoro-2-deoxy-D-glucose, R. L. Wahl, R. L. Cody, G. D. Hutchins, E. E. Mudgett, Radiology (1991) 179:765-770). PET imaging of breast cancer patients given FDG has been shown to be useful in imaging tumors as small as 3.2 cm and in patients whose breasts are too dense to be imaged well mammographically (Primary and Metastatic Breast Carcinoma: Initial Clinical Evaluation with PET with the Radiolabeled Glucose Analogue 2-[F-18]-Fluoro-2-deoxy-D-glucose, R. L. Wahl, R. L. Cody, G. D. Hutchins, E. E. Mudgett, Radiology (1991) 179:765-770).
The use of a specially collimated planar gamma camera to image the breast with this high resolution is limited by technical factors. The energy of 511 KeV is too penetrating to be detected well by conventional gamma cameras, and the collimation required to correct for the high energy leads to loss of signal (counts/pixel) that is equivalent to resolution loss due to low photon flux.
Conventional PET imaging devices are designed to image cross sections of the entire body. Accordingly, there are several disadvantages to employing a whole body PET scanner in a primary role as a high resolution confirmatory modality for small suspicious breast lesions. The first disadvantage of using a whole body PET scanner for breast imaging is the limited resolution available. The net resolution of a whole-body PET system is a combination of individual factors and is limited to above 5 mm FWHM (E. Rota-Kops et al., Journal of Computer Assisted Tomography 1990, May-June 14 (3), pages 437-445; N. A. Mullani et al., Journal of Nuclear Medicine 1990, May 31 (5), pages 610-616 and pages 628-631; K. Wienhard et al., Journal of Computer Assisted Tomography 1992, Sept.-Oct. 16 (5) pages 804-813). The effect of this resolution limit is that radioactivity is underestimated (Positron Emission Tomography and Autoradiography, Edited by M. E. Phelps, J. C. Mazziotta, H. R. Schelbert, pp. 240-285, Raven Press, N.Y. 1986; Design of a Mosaic BGO Detector System for Positron CT, H. Uchida, T. Yamashita, M. Iida, S. Muramatsu, IEEE Transactions on Nuclear Science February 1986, NS-33 (1), pp. 464-467). This reduces the sensitivity of PET scanners in estimating the malignancy of mammographically detected lesions smaller than twice the resolution limit, and also precludes the use of the PET scanner in delineating tumor margins with high accuracy.
A second disadvantage of a conventional PET scanner for imaging of subtle lesions in the breast is the high cost of the examination. In order to accommodate the entire body, a conventional PET scanner must employ tens or hundreds of expensive detector arrays along with a gantry and associated electronics.
A third disadvantage of a PET scanner is that the PET image format would not be easily compared to conventional mammograms. This is due to the fact that the breast is an organ which can be compressed to an essentially two-dimensional object. The variability in internal architecture of the breast results in few landmarks for positioning, and the location of an anomaly on the mammographic image of the compressed breast does not always correspond to the same location in the non-compressed breast.
In order to achieve the highest spatial resolution available in a tomographic system, motion of the patient due to breathing must be limited. Immobilizing of the breast by compression is the most straightforward approach to solving this problem, but implementation within a PET scanner detector ring is difficult. Additionally, the use of PET scanner to image an essentially two-dimensional object such as a compressed breast is not economically rational.
High resolution (20 cm diameter bore) PET scanners, originally developed for animal studies, may soon be available commercially. For a system with smaller aperture (i.e. 20 cm bore for a dedicated head scanner) the resolution in the axial plane is 3.5 mm (Development of a High Resolution PET, T. Yamashita et al., IEEE Transactions on Nuclear Science, April 1990, Vol. 37 (2) pp. 594-599). Such a system would satisfy the goal of high resolution. A disadvantage would be the considerable cost of such relatively expensive scanners, with approximately fifteen detector arrays, as dedicated units for breast imaging. Further, the problems of immobilization of the breast and of comparison to standard mammography would still be unaddressed.